A. Khawaja, and O. Dössel. Predicting the QRS complex and detecting small changes using principal component analysis. In Biomed Tech (Berl), vol. 52(1) , pp. 11-17, 2007
In this paper, a new method for QRS complex analysis and estimation based on principal component analysis (PCA) and polynomial fitting techniques is presented. Multi-channel ECG signals were recorded and QRS complexes were obtained from every channel and aligned perfectly in matrices. For every channel, the covariance matrix was calculated from the QRS complex data matrix of many heartbeats. Then the corresponding eigenvectors and eigenvalues were calculated and reconstruction parameter vectors were computed by expansion of every beat in terms of the principal eigenvectors. These parameter vectors show short-term fluctuations that have to be discriminated from abrupt changes or long-term trends that might indicate diseases. For this purpose, first-order poly-fit methods were applied to the elements of the reconstruction parameter vectors. In healthy volunteers, subsequent QRS complexes were estimated by calculating the corresponding reconstruction parameter vectors derived from these functions. The similarity, absolute error and RMS error between the original and predicted QRS complexes were measured. Based on this work, thresholds can be defined for changes in the parameter vectors that indicate diseases.
Following the ICH E14 clinical evaluation guideline , the measurement of QT/QTc interval prolongation has become the standard surrogate biomarker for cardiac drug safety assessment and the faith of a drug development. In Thorough QT (TQT) study, a so-called positive control is employed to assess the ability of this study to detect the endpoint of interest, i.e. the QT prolongation by about five milliseconds. In other words the lower bound of the one- sided 95% confidence interval (CI) must be above 0 [ms]. Fully automated detection of ECG fiducial points and mea- surement of the corresponding intervals including QT in- tervals and RR intervals vary between different computer- ized algorithms. In this work we demonstrate the ability and reliability of Hannover ECG System (HES) to as- sess drug effects by detecting QT/QTc prolongation effects that meet the threshold of regulatory concern as mentioned by using THEW database studies namely TQT studies one and two.
A. Khawaja, and O. Dössel. A PCA-Based Technique for QRS Complex Estimation. In Proc. Computers in Cardiology, vol. 32, pp. 747-750, 2005
In this paper, a new method for QRS complex prediction is presented. It is based on Principal Components Analysis (PCA) and polynomial fitting techniques. QRS complexes were extracted from multi-lead ECG signals and were aligned very perfectly. The covariance matrix was calculated from the QRS complex data matrix of many heartbeats. Afterwards, the corresponding eigenvectors and eigenvalues were computed and the reconstruction parameters vectors were derived by expansion of every beat in terms of the first eigenvectors. Performing the first order poly-fit method on the elements of the reconstruction parameter vectors yielded certain linear functions. Thereafter, the following QRS complexes were estimated by calculating the corresponding reconstruction parameter vectors derived from these functions. The similarity, absolute error and RMS error between the original and predicted QRS complexes were measured
Detecting peaks and boundaries of ECG characteristic waves supplies fundamental features for extracting clinically useful information. In this paper, an accurate threshold-independent multi-lead ECG delineation system is presented. Detection of QRS complexes, P and T waves is based on wavelet transform using Haar function as prototype wavelet and analyzing the first scale details coefficients.The delineator is performed on certain selected channels. Afterwards, a method, using a special histogram-based estimation, yields the exact positions of the significant points in all multi-channel ECG signals. The algorithm is applied on MIT-BIH Arrhythmia database signals and on multi-channel ECG signals measured at our institute.The single-channel delineation method was tested on MIT-BIH Arrythmia database signals. Sensitivity and positive predictivity were greater than 99.84% and 99.89% respetively for more than 15,990 beats. Furthermore, an overall mean error of less than two sampling intervals (1 ms) is obtained comparing the manual and the automatic method, whereas the standard deviation does not exceed three sampling intervals.
A. Khawaja, S. Sanyal, and O. Dössel. A wavelet-based technique for baseline wander correction in ECG and multi-channel ECG. In IFMBE Proceedings, vol. 9, pp. 291-292, 2005
In this paper, a new offline method for automatic baseline drift correction in Electrocardiogram is presented. It is based on Discrete Wavelet Transform (DWT) and analyzing high scale Approximation Coefficients (AC). A set of 650 noisy ECG signals was created by mixing different artificially generated noise-free ECGs and baseline wanders. By applying different mother Wavelets on each noisy signal, twelve stage DWT decomposition was carried out and twelve filtered ECGs were reconstructed by canceling the highest level AC at each stage. The similarities between initially generated baseline and canceled AC, as well as between the corresponding noise-free and reconstructed ECGs were examined every time by means of Correlation technique. The results from all 650 signals were considered in order to find the suitable Wavelet and AC level. The highest correlations, better than 99.9% for baseline and 99.99% for filtered ECG, were found with the ninth scale approximation coefficients when using Daubechies11 or Symlet12 as prototype wavelet. The algorithm was applied on various MIT-BIH and Multi-channel ECG signals. Furthermore, the baseline elimination results were considered to be very promising.
T. Baas, K. Gräfe, A. Khawaja, and O. Dössel. Investigation of parameters highlighting drug induced small changes of the T-wave's morphology for drug safety studies. In Conf Proc IEEE Eng Med Biol Soc, pp. 3796-3799, 2011
In guideline E14, the American Food and Drug Administration (FDA) requests for clinical studies to investigate the prolongation of the heart rate corrected QT-interval (QTc) of the ECG. As drug induced QT-prolongation can be caused by changes in the repolarisation of the ventricles, it is so far a thorough ECG biomarker of risk for ventricular tachyarrhythmias and Torsade de Pointes (TdP). Ventricular repolarisation changes not only change QT but also influence the morphology of the T-wave. In a (400 mg single dose) Moxifloxacin positive control study both, QTc and several descriptors describing the T-wave morphology have been measured. Moxifloxacin is changing two shape dependent descriptors significantly (P<0.05) about 3 to 4 hours after a 400 mg oral single dose of Moxifloxacin.
There is a large interest in analysing the QT-interval, as a prolonged QT-interval can cause the development of ventricular tachyarrhythmias such as Torsade de Pointes. One major part of QT-analysis is T-end detection. Three automatic T-end delineation methods based on wavelet fil- terbanks (WAM), correlation (CORM) and Principal Com- ponent Analysis PCA (PCAM) have been developed and applied to Physionet QT database. All algorithms tested on Physionet QT database showed good results, while PCAM produced better results than WAM and CORM achieved best results. Standard de- viation in sampling points (fs=250Hz) have been 33.3 (WAM), 8.0 (PTDM) and 7.8 (CORM). It could be shown that WAM is prone to interference while CORM is the most stable method even under bad conditions. Further- more it was possible to detect significant QT-prolongation caused by Moxifloxacin in Thorough QT Study # 2 us- ing CORM. QT-prolongation is significantly correlated to blood plasma concentration of Moxifloxacin.
S. Seitz, A. Khawaja, and O. Dössel. PCA-based method for clustering T-waves. In Gemeinsame Jahrestagung der Deutschen, der Österreichischen und der Schweizerischen Gesellschaft für Biomedizinische Technik, 2006
A. Khawaja. Automatic ECG analysis using principal component analysis and wavelet transformation. Universitätsverlag Karlsruhe. Dissertation. 2006
The main objective of this book is to analyse and detect small changes in ECG waves and complexes that indicate cardiac diseases and disorders. Detecting predisposition to Torsade de Points (TDP) by analysing the beat-to-beat variability in T wave morphology is the main core of this work. The second main topic is detecting small changes in QRS complex and predicting future QRS complexes of patients. Moreover, the last main topic is clustering similar ECG components in different groups.